CN1755669A - Name input processing method and system - Google Patents

Abstract

The invention discloses an input method and equipment which can identify the inputted name. The method comprises: checking the surname inputting code in the inputting coded sequence, choosing the candidate character which is corresponding with the surname inputting code in surname listing, acquiring the name's word number from the inputting coded sequence, if the number of the name's word is a single name, calculating the using probability of the single candidate word and doing descend arrangement by the using probability; if the number of the name's word is a double name, calculating the suing probability of the double candidate word and doing descend arrangement by the using probability of the candidate word.

Description

Name input processing method and system

technical field

The present invention relates to a method and system for inputting Chinese character names, in particular to a method and system capable of rapidly inputting Chinese names, reducing character selection and realizing continuous input.

Background technique

Unlike the alphabetic scripts of Indo-European languages, Oriental languages, such as Chinese, Japanese, and Korean, often have a very large character set. The number of Chinese characters in GB2312 reaches 6763, while the number of Chinese characters in GBK exceeds 20,000, so the input of Chinese characters is always considered the first problem of Chinese informatization. The same problem also exists in Japanese and Korean input. In order to solve the problem of Chinese character input, many input methods have been born in the past 20 years, including the famous Wubi font and intelligent ABC, etc. Ziguang Pinyin input method and Microsoft Pinyin input method also solve the problem of reducing word selection and continuous input to a certain extent through the language statistical model.

Ordinary input methods convert from pinyin to Chinese characters through vocabulary lists. However, personal names are different from fixed entries and have the characteristics of dynamic combination. Therefore, it is impossible to pre-record a large enough list of personal names, which often requires users to input names. After a complicated multiple word selection.

Contents of the invention

The purpose of the present invention is to provide a method capable of efficiently inputting a person's name, which can be embedded in an existing input method, or can be designed as an independent input method. As a software system, it can be used for any kind of information equipment such as desktop computers, personal digital assistants and mobile phones that need to input names.

According to one aspect of the present invention, there is provided a kind of input method capable of identifying the input person's name, comprising the steps of: detecting the surname input code in the input code sequence; retrieving the candidate characters of the surname corresponding to the input surname input code from the surname list; Obtain the word count of the name from the input coding sequence; when it is judged that the word count of the name is a single name, then calculate the use probability of the single name candidate word as the target, and arrange the candidate words in descending order according to the use probability of the candidate word; And when it is judged that the number of characters in the name is a double name, the usage probability of the target double-name candidate is calculated, and the candidate characters are arranged in descending order according to the usage probability of the candidate word.

According to another aspect of the present invention, there is provided a kind of input method capable of identifying the input person's name, comprising the steps of: detecting the surname input code in the input code sequence; retrieving the candidate of the Chinese character surname corresponding to the input surname input code from the surname list word; input single-name syllable and look up the single-name index table stored; and utilize expression S(w)=0.5*(f(w, G ₀ |p)+f(w, G ₂ |p)) to obtain this pinyin All corresponding names use candidate characters.

According to another aspect of the present invention, there is provided a kind of input method capable of identifying the input person's name, comprising the steps of: detecting the surname input code in the input code sequence; retrieving the candidate of the Chinese character surname corresponding to the input surname input code from the surname list word; input a double name syllable and look up the stored double name index table; generate a double name candidate combination corresponding to the input code, and use the expression $S\left(w_1{w}_2\right)=\sqrt{f({\mathrm{<figure-callout\; id="1"\; label="w"\; filenames="A20041008318700181.png"\; state="\{\{state\}\}">w</figure-callout>}}_1,{\mathrm{<figure-callout\; id="1"\; label="G"\; filenames="A20041008318700181.png"\; state="\{\{state\}\}">G</figure-callout>}}_1|p)\&times;f({\mathrm{<figure-callout\; id="2"\; label="w"\; filenames="A20041008318700181.png"\; state="\{\{state\}\}">w</figure-callout>}}_2,{\mathrm{<figure-callout\; id="2"\; label="G"\; filenames="A20041008318700181.png"\; state="\{\{state\}\}">G</figure-callout>}}_2|p)}\&times;(1+f\left({\mathrm{<figure-callout\; id="2"\; label="w"\; filenames="A20041008318700181.png"\; state="\{\{state\}\}">w</figure-callout>}}_2\right|w_1)+f\left({\mathrm{the\; <figure-callout\; id="2"\; label="s"\; filenames="A20041008318700181.png"\; state="\{\{state\}\}">s</figure-callout>}}_2\right|{\mathrm{the\; s}}_1))$ evaluating each generated double name candidate; and outputting the double name candidate with the highest score as a double name candidate.

According to another aspect of the present invention, there is provided a kind of input method capable of identifying the input person's name, comprising the steps of: obtaining the current syllable from the currently input syllable sequence; querying the list of surnames to determine whether the list of surnames is empty; if the list of surnames is empty , then end the name input judgment, and return an empty result, if the surname candidate list is not empty, then use the first candidate in the surname list as the output surname, and list the surname candidate words; if the surname output is judged to be incorrect, Then select a surname candidate; if the output result is the required surname, then calculate the single-name candidate or the candidate double-name candidate from the input syllable sequence.

According to another aspect of the present invention, there is provided an input device capable of identifying the name of an input person, including: an input coding device for converting Chinese characters into an acceptable input coding sequence; a surname processing device for detecting and judging the input of the user The surname in the input code; the name processing device is used to identify the name in the subsequent pinyin sequence after the surname processing unit detects the surname in the input code; and the name output device is used to output the name candidate Chinese characters corresponding to the input code.

According to the method of the present invention, the name of a person in Chinese characters can be input efficiently, and a slight language-related modification of the method is also suitable for inputting Japanese and Korean.

Description of drawings

The above and other objects, features and advantages of the present invention will be more clearly described through the following detailed description of the preferred embodiments of the present invention in conjunction with the accompanying drawings. It should be pointed out that the description given below is only an example provided for a better understanding of the present invention, rather than limiting the present invention. in:

Fig. 1 is a schematic block diagram representing a name input system according to an embodiment of the present invention;

Fig. 2 is to represent the reverse index table of Chinese surname;

Fig. 3 shows the single name recognition and input of the personal name input system in the independent input mode according to the embodiment of the present invention;

Fig. 4 shows the double-name recognition and input of the personal name input system in the independent input mode according to an embodiment of the present invention; and

Fig. 5 shows an input interface with a name input function according to an embodiment of the present invention.

Detailed ways

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings, and unnecessary details and functions for the present invention will be omitted during the description to prevent confusion in the understanding of the present invention.

The basic principle of the input method according to the present invention is described below. It should be pointed out that although the Chinese character name input is described in the embodiment of the present invention, the idea and principle of the present invention can be applied to Chinese character input methods in other fields. In addition, the present invention can also be applied to Japanese and Korean input by making language-related modifications.

The present invention is described below by taking Chinese as an example. If the special case of names of ethnic minorities is not considered, names in Chinese usually have only four limited combination forms, namely, single surname and single name, single surname and double name, compound surname and single name, and compound surname and double name.

Because the combination possibility of surname and first name is very large, the common input method cannot provide a full name vocabulary. However, Chinese names and their combinations have certain meanings, rather than arbitrary combinations. In addition, Chinese surnames are limited in characters. Therefore, there are distribution characteristics in the sense of probability. The present invention utilizes this feature and uses a name recognition algorithm to help users reduce the number of word selections during the conversion process from codes to Chinese characters, so as to efficiently input names.

In order to clearly illustrate the problem, this embodiment takes pinyin input as an example to illustrate the personal name input method of the present invention. It should be pointed out that although pinyin input is used as an example in this embodiment, the present invention is not limited thereto, and the method proposed in the present invention is also applicable to other types of input methods, such as stroke-based input methods.

The personal name input method of the present invention includes two application modes, that is, an independent mode when the current input content is known to be a personal name, and a continuous mode when continuously inputting Chinese characters. In the stand-alone mode, the input method is set to only input Chinese names, after which the input string will be considered as a name, and a name recognition algorithm will be used to evaluate and select the candidate Chinese characters generated by the input pinyin, so that Users can complete input tasks with fewer word selections. The independent mode is suitable for occasions where a large number of names need to appear at the same time, for example, entering salary tables, personnel lists, etc. In the continuous mode, the name recognition algorithm dynamically detects the names of people that may exist in the pinyin string input by the user, and assists the user in word selection, thereby improving the efficiency of input. The continuous mode is suitable for inputting people's names in general article input and other occasions. In the independent mode, the user can actively set the input method program to the name input mode or switch to the name input method, or the name input method can be automatically activated by the operating system, web browser and other environments, such as filling in the name item in the web form. The continuous input mode is used to improve the efficiency of name input in large text input. In both modes, the present invention achieves the purpose of efficiently inputting names through a name recognition algorithm. The name recognition algorithm uses the set recognition method and mathematical model to sort the candidate characters generated by the input pinyin code, and finally generates the combination of Chinese characters with the highest probability, thereby reducing the number of times the user selects characters when inputting the name and achieving efficient input the goal of. The operation of the algorithm in the two modes is different, which will be described separately below.

The following describes the name recognition model and its establishment according to the present invention according to the principle of the present invention. Name Recognition Model and Its Establishment

The name recognition model is a mathematical model, which counts the necessary mathematical parameters from a pre-established name database, and is used to evaluate the possibility of a Chinese character (combination) as a Chinese name candidate in the input stage.

A series of real names are listed in the name database, for example: "Chen Geng, Zhang Zhizhong, Feng Yuxiang, Dai Anlan" and so on. For the four limited combinations of Chinese names, single and double names are processed separately. This is because from the perspective of mathematical statistics, the amount of information possessed by a single name and a double name is different.

For a single name, suppose we examine the situation that a Chinese character w is used as the single name G ₀ in a name. In order not to lose its generality, the word "Geng" can be used as an example. When the user inputs the pinyin "geng", in the database of personal names, the names with the pronunciation of "geng" include "geng, geng, geng, heng, geng, heng" and so on. These homonyms are counted and counted as the words used in names, and are arranged in descending order of their frequency of occurrence in names, so as to predict the words used in single names. In general, in order to predict single-name characters, it is necessary to establish an index from pinyin codes to homonym candidate single-name characters, and all single-name characters are sorted in descending order according to a specific frequency to form a single-name index table. As shown in Table 1 below:

The numbers in the table indicate how often the word appears in the name. The index table can be stored in a storage device of an information processing device such as a computer, that is, as a recognition model for predicting single-name characters. The mathematical description of the index table is: the frequency of all Chinese characters whose pinyin code is p, as a single name, is represented by the following expression (1).

f(w, G ₀ |p) = the number of times a single name word w is used...(1)

Likewise, for double names, Chinese characters w ₁ and w ₂ can be assumed as the first character G ₁ and the second character G ₂ of the double name respectively. In order not to lose generality, take the double name "Zhizhong" as an example. When the user inputs the pinyin "zhizhong", it is like using a single name, and the name database has lists of the same names in which zhi and zhong appear as the first and second characters in double names respectively. The occurrence frequency of the name "Zhizhong" is counted and an index table of words used for a single name is established. Their mathematical descriptions are:

f(w, G ₁ |p)=encoded as p double name use word w as the use times of the first word of double name

f(w, G ₂ |p)=coded as the number of times that the word w of the double name used as the second word of the double name is coded as p

…(2)

In addition, the connection frequency between two single characters of a double name is an effective method to measure the possibility of two independent Chinese characters forming a double name. The principle is that the binary connection parameters have a strong memory effect on the names that have appeared. Therefore, the word connection frequency f(w ₂ |w ₁ ) is introduced:

The above formula indicates that the single-character connection frequency is equal to the number of Chinese characters w ₁ and w ₂ as double names divided by the number of times Chinese character w ₁ is used as the first word in a double name. The number of times that Chinese characters w ₁ and w ₂ are used as double names includes the times that Chinese characters w ₁ and w ₂ appear in double names respectively and together.

In addition, as an alternative, the concatenation frequency of toned syllables using double names is also a better measurement method. The principle is that Chinese names attach importance to the smoothness of reading aloud. Therefore, the tonal syllable connection frequency f(s ₂ |s ₁ ) function is introduced:

Where s represents the syllable of the input word.

As shown in the expressions (1)-(4) above, the expression (1) counts the target single-name word frequency, and is used to describe the use frequency of a single-name word. This value can be used as the basis for sorting single names. Expression (2) counts the frequency of Chinese characters in different positions whose target is double-name characters. The statistical objectives of expressions (3) and (4) are used to evaluate the combination ability of two names with words and their pronunciation.

For a single name, considering that a single name is often used as the second word of a double name, the use score is set as:

E1: S(w)=0.5*(f(w, G ₀ |p)+f(w, G ₂ |p)) (5)

(Please elaborate on the meaning of expression (5))

For double names, assuming that a Chinese character candidate combination produced by two pinyins is w ₁ w ₂ , its usage score is set as:

E2: $S\left(w_1{w}_2\right)=\sqrt{f({\mathrm{<figure-callout\; id="1"\; label="w"\; filenames="A20041008318700181.png"\; state="\{\{state\}\}">w</figure-callout>}}_1,{\mathrm{<figure-callout\; id="1"\; label="G"\; filenames="A20041008318700181.png"\; state="\{\{state\}\}">G</figure-callout>}}_1|p)\&times;f({\mathrm{<figure-callout\; id="2"\; label="w"\; filenames="A20041008318700181.png"\; state="\{\{state\}\}">w</figure-callout>}}_2,{\mathrm{<figure-callout\; id="2"\; label="G"\; filenames="A20041008318700181.png"\; state="\{\{state\}\}">G</figure-callout>}}_2|p)}\&times;(1+f\left({\mathrm{<figure-callout\; id="2"\; label="w"\; filenames="A20041008318700181.png"\; state="\{\{state\}\}">w</figure-callout>}}_2\right|w_1)+f\left({\mathrm{the\; <figure-callout\; id="2"\; label="s"\; filenames="A20041008318700181.png"\; state="\{\{state\}\}">s</figure-callout>}}_2\right|{\mathrm{the\; s}}_1))---\left(6\right)$

(Please explain in detail the meaning of the above formula)

name input system

The personal name input system of the present invention will be described below with reference to the accompanying drawings.

The input method with the function of efficiently inputting personal names according to the present invention can be used as a supplementary part of the existing input method, or used as independent software in any device that needs to input personal names. FIG. 1 is a structural block diagram of a name input system according to an embodiment of the present invention.

The personal name input system according to the present invention includes a code input device 1, a surname processing unit 2, a name processing unit 3, a personal name output device 4, a surname list storage unit 5, a recognition rule storage unit 6, and a recognition model storage unit 7.

The input encoding device 1 accepts the character encoding sequence input by the user, and provides the input character sequence to the surname processing unit 2 . The surname processing unit 2 uses the surname list storage unit 5 and the recognition rule storage unit 6 to perform surname detection, and provides the detected surname to the name processing unit 3 . The name processing unit 3 uses the recognition model 7 to perform name detection, outputs the detection result to the name output device 4, and the name output device 4 displays the name corresponding to the input character code sequence. The operation of each part in Fig. 1 will be described below.

code input

In order to input Chinese characters, users need to convert Chinese characters into computer-acceptable coded strings with some encoding method. For the input method based on Chinese Pinyin, its encoding string is the pinyin string of Chinese characters. Considering the most general Quanpin input, taking the input of the name "Xia Hairong" as an example, the user may input "xia hai rong". After the pinyin string is input into the system, it is parsed into three independent syllables "xia", "hai" and "rong".

Surname processing

Chinese has fairly stable surnames, and by creating a list of surnames (see Table 1), it can be used to detect and judge the surnames in the Pinyin input by the user. The surnames included in the surname list can be retrieved by encoding, for example, the surname "Xia" in the above example can be retrieved by inputting "xia". For another example, the surname "Ouyang" can be retrieved by entering "ou yang". The surname list may be stored in the surname list storage unit 5 .

Table 1: List of Chinese surnames (partial) surname Pinyin coding surname Pinyin coding Zhao zhao money money open zhang chapter Zhang Xiahou xia hou summer Xia Ouyang ou yang Europe Ou … … … …

When the table is in use, it is often necessary to establish a reverse index from code to surname, and its structure is shown in Figure 2. Through this reverse index table, it can be judged whether an input syllable may constitute a surname in a person's name, and how many homophonic surnames there are. For example, when the pinyin "ai" is input, the surname processing unit 2 can search the surname list storage unit 5 for the Chinese surname corresponding to the pinyin code "ai". Usually, the Chinese surname corresponding to the pinyin code "ai" is only "AI", therefore, the detected surname is both "AI".

There are a small number of compound surnames in Chinese surnames, such as "Xiahou, Ouyang, Yuchi, Zhuge", etc., and the first character itself can also be used as a single surname. In this case, as an example, the system can default it to form a compound name, but allow the user to modify it. That is, when the user inputs the pinyin "ouyang...", the surname processing unit 2 may consider that the surname of the person is "Ouyang" instead of "OU". And when the pinyin "ou" is input, the surname processing unit 2 can detect the characters "Europe" and "District" for the single surname, and as a setting, can detect the double surname "Ouyang".

In addition, Chinese surnames also include some homonyms, such as "Zhang" and "Zhang" listed in Figure 2. There is no obvious effective way to solve this problem, and the user can determine it through the word selection process.

For the user's input, if the system cannot find a suitable surname, the process of inputting the name of the person is stopped, and the normal character selection process of the input method is switched to the character selection input.

The above identification rules may be stored in the identification rule storage unit 6 .

name manipulation

Compared with the surname, the first name is much more extensive than the last name, so the main algorithm process described above is involved here. The procedure differs slightly depending on whether the system is in stand-alone or continuous mode. For a determined input method program, a special key combination or other control methods such as buttons and menu items can be set on the information processing device to switch the mode of name input.

1. Independent mode

In the stand-alone mode, the purpose of the user is to input a person's name, so the input code string is considered to be an effective name code. After the user's input is finished, the system can directly know the total number of characters of the name. After processing by the surname processing unit 2, the name processing unit 3 can know the number of characters of the name, that is, whether it is a single name or a double name. Therefore, the name processing unit 3 can independently process single names and double names.

Fig. 3 shows the operation flow of the system of the present invention to process single-name characters. First, in step 301, a single-named syllable s is input. Thereafter, in step S302, the name processing unit 3 searches the single name index table. Next, in step S303, use the expression (5) to obtain all the candidate characters for the name corresponding to the pinyin, and in step S304, select the single-name candidate character with the highest frequency as the output character. As an example, the singular name index table may be stored in the recognition model unit 7, or a separate storage unit may be provided for storing the name word index table. For example, it is divided into single-name index table and double-name index table.

Let’s take the user’s input of the name “chen geng” as an example to illustrate in detail. The surname processing unit first judges that “chen” is a single surname “Chen”, and then finds out the candidate characters of “geng” sorted by frequency according to the index table. ...", take the first candidate word "Geng" as the selection output, and keep the list for the user to choose.

Fig. 4 has shown the operation flow of the system of the present invention to handle double-named characters. First, in step 401, double name syllables s1, s2 are input. Thereafter, in step S402, the name processing unit 4 searches the stored dual name index table. Next, generate double-name candidate combinations corresponding to the input pinyin in step S403, and evaluate each double-name candidate combination generated in step S403 by using expression (6) in step S404. The double-name candidate with the highest score is output as the word for output in step S405. The output order can be arranged according to the frequency of double-name combinations.

The following takes the user's input of "xia hai rong" as an example to describe in detail the process of identifying double names by the name input system in the stand-alone mode. The system first judges that "xia" is a single surname "Xia", and then finds the candidate characters of "hai" and "rong" sorted by frequency according to the double name index table, "hai, hai, return" and "rong, rong, rong, Melt, melt, melt". After combining and scoring with the expression (6), the double-name candidate words are sorted according to the score as follows: Hairong, Hairong, Hairong, Hairong, and the user can choose the candidate words given by the system. Finally, the system will use "Hairong" as the recognition output.

2. Continuous mode

The following describes the case of processing the input of a person's name in the continuous mode according to the present invention. In the continuous mode, the user continuously inputs the Chinese character encoding string, and there is no obvious mark to mark the beginning and end of the name, so the system needs to dynamically detect the name of the person in the Chinese character encoding string input by the user. Its basic algorithm is as follows:

First, the current syllable variable s and the candidate word w of s can be set. Get the current syllable s from the currently input syllable sequence. Thereafter, the surname list is queried to obtain a surname candidate list 1 . Check if the list of surnames is empty. If the last name list is empty, end the name input judgment in continuous mode and return an empty result. If the surname candidate list is not empty, the first candidate in the surname list is used as the output surname. Thereafter, the user can judge whether the output surname is correct, and if the input is incorrect, that is, the output surname is not the surname required by the user, the user can select a candidate character. If the output is a character other than a surname, the surname recognition process ends, an empty result is returned, and Chinese character input is continued. If the output result is the surname required by the user, it is assumed that the user has entered a single name and a double name respectively. Thereafter, according to the above expressions (5) and (6), score calculations are performed on the input coded pinyin, and the single-name and double-name characters with the highest scores are displayed to the user.

Thereafter, the user judges whether the selected person's name is correct. If it is correct, the current recognition of the person's name input is ended. If the user judges that the selected person's name is incorrect, then select the first word of the name according to the given candidate words. If the user determines that the input of the person's name is finished, the current recognition of the person's name input is ended. If the user determines that the name input has not ended, the system of the present invention performs score calculation according to expression (6), and selects the double name with the highest score to be displayed to the user. The user can judge whether the second character of the selected person's name is correct. If it is correct, the current name input is ended. If it is not correct, then select the second word of the name to end the identification of this person's name input.

For clarity, the above process can be represented by the following table: Algorithm: Detection of personal names in continuous mode Variable: current syllable s, candidate word w for s 1. Obtain the current syllable s from the input syllable sequence 2. Query the surname list and obtain a surname candidate list l3. If the surname candidate list l is empty 3.1 end, return an empty result 4. Otherwise 4.1 take the first candidate in the list as the output surname 4.2 The user judges whether the surname is correct, if not, then 4.2.1 The user selects a candidate word, if it is a non-surname character, then 4.2.1.1 ends and returns an empty result 4.2.2 Otherwise 4.2.2.1 Assume single name and double name respectively 4.2 .2.2 Scoring according to the formulas E1 and E2 respectively 4.2.2.3 Display the single name and double name with the highest score to the user 5. The user judges whether the selected name is correct, if it is correct 5.1 End this name recognition input 6. Otherwise 6.1 User Select the first character of the name 6.2 If the user determines that the name is over, then 6.2.1 End this name recognition input 6.3 Otherwise 6.3.1 Score according to E2, select the double name with the highest score and display it to the user 6.4 The user judges whether the second character of the selected person’s name is correct, if correct 6.4.1 End this name recognition input 6.5 Otherwise 6.5.1 User selection The second character of the name, ending this name recognition input

The following is an example of the recognition process of the appointment input method in the continuous mode according to the present invention. For example, suppose the user inputs the pinyin string "wo he li bao ying lao shi you guo yi mianzhi yuan", and the target string is "I had a relationship with Teacher Li Baoying once". Find the corresponding surname for each syllable, the list is as follows: last name syllable Surname candidate single name hypothesis double name hypothesis the w fertile he he li he He, He, He, He, He li li bao li Li, Li, Li, Li bao bao ying bao bag, abalone ying ying lao ying Ying, should, win lao lao shi lao Law, mash the sh shi you the sh stone, teacher, Shi, history, time you you guo you especially guo guo yi guo Guo, fruit, country yi yi mian yi Yi, Yi, Yi mian mian zhi mian the zhi yuan Yuan, Yuan, Yuan

According to the identification process of the present invention, after the system first converts "wo" into the hypothetical surname "wo", the user selects "I" as a non-surname in 4.2.1, returns empty, and the input of this person's name recognition ends. Execute to "li" in turn, the user selects "Li" in 4.2.1, and then the system assumes a single name and a double name, and recommends a candidate name "Li Baoying" to the user through calculation, and the user judges that the name recommended by the system is wrong in step 6.1 , so the user selects "Baoying" in 6.1, and then the system outputs the name "Li Baoying", which ends correctly. The system processes each syllable in turn according to the algorithm until the syllable string ends.

As an alternative, it is possible to record and count the user's inputted personal name after each input of the user's surname and personal name. If it is the same as the previously input surname or person's name, then use the above expressions (1)-(6) to update the input frequency of the corresponding person's name. Therefore, according to the frequency of use of the name, the ranking of candidates can be changed in real time for the pinyin code input by the user, so as to input the name more efficiently.

The input method running on the PC platform can usually have an input area and an information area, the user inputs codes such as Pinyin in the input area, and the Chinese character candidates corresponding to the codes are displayed in the information area. If the present invention works in an independent mode or as an independent input method, both the input area and the information area can be used directly, and the use process is similar to an ordinary input method. If working in continuous mode, in order not to affect the normal input and word selection process of the input method, a second information area for displaying the detection result of the person name can be added near the input area, for example, below the normal input method, as shown in Figure 6 shown.

During the normal process of inputting Chinese characters, the user inputs the Chinese character code through the first input area, and selects characters from the first information area to correct wrong Chinese characters in the input area. When the present invention detects a possible name, it draws the second information area on the screen, and the user can switch to the second information area through an undefined key of the input method, such as an arrow key or a Tab key, or use a mouse. Then select the candidate name provided by the present invention from the second information area.

The personal name input method of the present invention can be applied to devices that need to input Chinese characters, for example, personal computers (PCs), portable computers, mobile phones, PDAs (Personal Digital Assistants) and other devices.

The personal name input system according to the present invention can be realized by hardware. It can also be realized by using software or a combination of hardware and software. The program can be recorded on a machine-readable recording medium such as a floppy disk, a hard disk, a flash disk, a CD-ROM, and a DVD-ROM.

Although the invention has been described with reference to preferred embodiments, the invention is not limited thereto but only by the appended claims, and various modifications can be made to the embodiments of the invention by those skilled in the art without departing from the spirit of the invention. changes and improvements.

Claims (25)

1. can discern the input method of importing name for one kind, comprise step:

Detect the surname input coding in the input coding sequence;

From the surname tabulation, retrieve the time selected character of the surname corresponding with the surname input coding of input;

From the input coding sequence, obtain the number of words of name;

When the number of words of judging name is single-character given name, then calculates the probability of use of waiting selected character as the single-character given name of target, and, wait selected character by descending sort according to the probability of use of waiting selected character; With

When the number of words of judging name is two-character given name, then calculates the probability of use of waiting selected character as the two-character given name of target, and, wait selected character by descending sort according to the probability of use of waiting selected character.

2. method according to claim 1 is wherein calculated single-character given name and is waited the probability of use that the step of the probability of use of selected character is utilized following expression (1) calculating single-character given name time selected character,

F (w, G ₀| p)=the single-character given name access times of word w ... (1)

Wherein w represents Chinese character, G ₀The position of expression Chinese character in name, p represents that Pinyin coding is all Chinese characters of p, as the frequency of single-character given name with word.

3. method according to claim 1 is wherein calculated the step of waiting the probability of use of selected character as the two-character given name of target and is utilized following expression (2) to calculate the probability of use that two-character given name is waited selected character,

F (w, G ₁| p)=two-character given name that is encoded to p is with the access times of word w as two-character given name first word

F (w, G ₂| p)=two-character given name that is encoded to p is with the access times of word w as two-character given name second word

……(2)

W wherein ₁And w ₂The expression Chinese character, G ₁Expression two-character given name first word, G ₂Expression two-character given name second word, p represents that Pinyin coding is the access times of the two-character given name Chinese character of p.

4. method according to claim 3 is wherein calculated the step of waiting the probability of use of selected character as the two-character given name of target and is also comprised and utilize following expression (3) to calculate the step that two-character given name is waited the connection frequency of selected character,

F (w wherein ₂| w ₁) expression individual character connection frequency.

5. method according to claim 3 is wherein calculated the step of waiting the probability of use of selected character as the two-character given name of target and is also comprised and utilize following expression (4) to calculate the step that two-character given name is waited the connection frequency of selected character,

F (s wherein ₂| s ₁) expression band tuning joint connection frequency.

6. method according to claim 2 is wherein calculated single-character given name and is waited the score that the step of the probability of use of selected character is utilized the probability of use of following expression (5) calculating single-character given name time selected character,

S(w)＝0.5*(f(w，G ₀|p)+f(w，G ₂|p))……(5)

Described keep the score considered single-character given name with word also usually as second word of two-character given name with word.

7. method according to claim 3 is wherein calculated the step of waiting the probability of use of selected character as the two-character given name of target and is also comprised and utilize following expression (4) to calculate the score that two-character given name is waited the combination of selected character,

$S\left(w_1{w}_2\right)=\sqrt{f(w_1,G_1|p)\&times;f(w_2,G_2|p)}\&times;(1+f\left(w_2\right|w_1)+f\left(s_1\right|s_2))......\left(6\right)$

Described keeping the score considered two Chinese character candidate combinations w that input coding produces ₁w ₂

8. according to the described method of claim 1 to 7, wherein said input coding is a Chinese-character sound dissection encode.

9. can discern the input method of importing name for one kind, comprise step:

Detect the surname input coding in the input coding sequence;

From the surname tabulation, retrieve the time selected character of the Chinese character surname corresponding with the surname input coding of input;

Input single-character given name syllable is also searched the single-character given name concordance list of storage; With

Utilize expression formula S (w)=0.5* (f (w, G ₀| p)+f (w, G ₂| p)) obtain all name candidate of this phonetic correspondence.

10. method according to claim 9, wherein said input coding is a Chinese-character sound dissection encode.

11. the input method that can discern the input name comprises step:

Detect the surname input coding in the input coding sequence;

From the surname tabulation, retrieve the time selected character of the Chinese character surname corresponding with the surname input coding of input;

Input two-character given name syllable is also searched the two-character given name concordance list of storage;

Generate and the corresponding two-character given name candidate combinations of importing of input coding, and utilize expression formula $S\left(w_1{w}_2\right)=\sqrt{f(w_1,G_1|p)\&times;f(w_2,G_2|p)}\&times;(1+f\left(w_2\right|w_1)+f\left(s_1\right|s_2))$ Estimate each the two-character given name candidate combination that generates; With

Output has the two-character given name candidate combinations of top score and waits selected character as two-character given name.

12. method according to claim 11, wherein said input coding is a Chinese-character sound dissection encode.

13. the input method that can discern the input name comprises step:

From the syllable sequence of current input, obtain current syllable;

The tabulation of inquiry surname judges whether the surname tabulation is empty;

Judge if surname tabulation, then finishes the name input for empty, and return the sky result, if the surname candidate list be empty, then with first candidate in the surname tabulation as the output surname, and list surname time selected character;

If judge that the surname of output is incorrect, then select a surname to wait selected character;

If the output result is the surname that needs, then from the syllable sequence of input, calculates single-character given name and wait selected character or candidate's two-character given name word.

14. method according to claim 13 wherein further comprises (f (w, G according to expression formula S (w)=0.5* ₀| p)+f (w, G ₂| p)) calculate the score that single-character given name is waited selected character, and show the step of single-character given name time selected character with top score.

15. method according to claim 13 wherein further comprises according to expression formula $S\left(w_1{w}_2\right)=\sqrt{f(w_1,G_1|p)\&times;f(w_2,G_2|p)}\&times;(1+f\left(w_2\right|w_1)+f\left(s_2\right|s_1))$ Calculate two-character given name and wait the score of selected character, and show the step of single-character given name time selected character with top score.

16. method according to claim 13 wherein further comprises when judging that the name of selecting is incorrect, then selects the step of first word of name according to the candidate that provides.

17. method according to claim 16, wherein further being included in the name input does not have under the situation of end, according to expression formula $S\left(w_1{w}_2\right)=\sqrt{f(w_1,G_1|p)\&times;f(w_2,G_2|p)}\&times;(1+f\left(w_2\right|w_1)+f\left(s_2\right|s_1))$ Carry out score calculation, and the two-character given name of selecting to have top score is shown to user's step.

18. method according to claim 17 comprises further wherein whether second word of the name of judge selecting be correct, if incorrect, then selects second word of name, finishes the step of name input identification.

19. method according to claim 13 wherein further comprises if what export is non-surname word, then finishes the surname identifying, continues the Chinese character input.

20. according to the described method of claim 13-19, the phonetic transcriptions of Chinese characters of the syllable of wherein said input.

21. the input media that can discern the input name comprises:

The input coding device is used for converting Chinese character to acceptable input coding sequence;

The surname treating apparatus is used for detecting and judging the surname of the input coding that the user imports;

The name treating apparatus is used for discerning the name in the follow-up pinyin sequence after the surname processing unit detects the surname of input coding; With

The name output unit is used to export the name candidate Chinese character corresponding with input coding.

22. device according to claim 21 wherein further comprises surname list storage device, is used to store the tabulation of Chinese character surname, described surname can be retrieved by coding.

23. device according to claim 21 comprises further that wherein it still is the two-character surname surname that the recognition rule storage unit is used to store the corresponding monosyllabic name of being imported according to the input coding judgement of input of phonetic.

24. device according to claim 21, wherein further comprise the model of cognition memory storage, be used to store the mathematical model of carrying out name identification, described mathematical model is the necessary mathematic parameter of statistics from the database of names of setting up in advance, is used for estimating the probability of the combination of a Chinese character or Chinese character as Chinese name candidate Chinese character at input phase.

25. according to the described device of claim 21 to 24, wherein said input coding is a Chinese-character sound dissection encode.

Images